The Influence of Salinity on the Diet of Nesting Bald Eagles

J. Raptor Res. 42(2):99–109 E 2008 The Raptor Research Foundation, Inc.

THE INFLUENCE OF SALINITY ON THE DIET OF NESTING BALD EAGLES

A. CATHERINE MARKHAM1 AND BRYAN D. WATTS Center for Conservation Biology, College of William and Mary, P.O. Box 8795, Williamsburg, VA 23187 U.S.A.

ABSTRACT.—Although the breeding density of Bald Eagles (Haliaeetus leucocephalus) in the lower Chesapeake Bay is known to vary with salinity, the ecological factors that contribute to this distribution have not been explored. In an effort to examine whether variation in prey use is associated with nest density patterns, we investigated the influence of salinity (tidal-fresh vs. mesohaline zones) on Bald Eagle diet composition by using video-monitoring to observe food delivered to nests during the 2002–2003 breeding seasons. Deliv- ered prey items were identified to the lowest taxonomic level possible and sizes were estimated relative to eagle bill length. We used species-specific length-weight relationships for prey to estimate biomass delivery. Overall, the diet included at least 12 species of fishes, three species of birds, four species of mammals, and four species of reptiles. Salinity had no significant influence on diet composition; Ictaluridae and Clupei- dae species were the most frequent prey items in both salinity zones. We suggest that pairs nesting in both tidal-fresh and mesohaline zones have access to similar fish species. However, the length and biomass of fish prey varied with salinity such that larger prey on average were delivered to nests in the mesohaline reaches compared to tidal-fresh zones. Thus, we suggest that foraging eagles may be exploiting more energetically- favorable conditions in higher salinity waters. Temporally, diet composition varied between study years, potentially reflecting annual changes in the availability of prey species. We consider differences in weather patterns between study years as the most likely factor contributing to this interannual variation in diet.

KEY WORDS: Bald Eagle; Haliaeetus leucocephalus; anadromous; Chesapeake Bay; diet; salinity.

INFLUENCIA DE LA SALINIDAD SOBRE LA DIETA DE HALIAEETUS LEUCOCEPHALUS DURANTE LA ANIDACIO´ N

RESUMEN.—Aunque se sabe que la densidad de aves reproductivas de la especie Haliaeetus leucocephalus en la parte baja de la bahıá de Chesapeake varıá con la salinidad, los factores ecolo´gicos que contribuyen a esta distribucioń no han sido explorados. En un esfuerzo para examinar si la variacioń en el uso de presas esta´ asociada con los patrones de densidad de nidos, investigamos la influencia de la salinidad (zona mareal- dulce vs. zona mesohalina) sobre la composicioń de la dieta de H. leucocephalus. Para ello utilizamos ca´maras de video para monitorear el alimento llevado a los nidos durante las e´pocas reproductivas de 2002 y 2003. Los´tems ı de presas llevados fueron identificados al menor nivel taxono´mico posible y sus tamanõs fueron estimados con relacioń a la longitud del pico de las a´guilas. Utilizamos relaciones de longitud-peso especı´ficas para las especies de presa para estimar la biomasa de alimentos entregada. En general, la dieta incluyo´ al menos 12 especies de peces, tres de aves, cuatro de mamı´feros y cuatro de reptiles. La salinidad no tuvo una influencia significativa sobre la composicioń de la dieta; las especies de Ictaluridae y Clupeidae fueron los´tems ı ma´s frecuentes entre las presas en ambas zonas de salinidad. Sugerimos que las parejas que anidan tanto en las zonas mareales-dulces como en la mesohalinas tienen acceso a especies de peces similares. Sin embargo, la longitud y la biomasa de peces presa vario´ con la salinidad de modo que, en promedio, se entregaron presas ma´s grandes en los nidos de las a´reas mesohalinas que en los de las a´reas mareales-dulces. Por lo tanto, sugerimos que las a´guilas podrıán estar explotando condiciones energe´ticas ma´s favorables al buscar alimento en aguas con mayor salinidad. Temporalmente, la composicioń de la dieta vario´ entre los anõs de estudio, lo que potencialmente refleja cambios anuales en la disponibilidad de especies de presas. Consideramos que las diferencias en los patrones clima´ticos entre los anõs de estudio son el factor que ma´s probablemente contribuye a la variacioń interanual en la dieta. [Traduccioń del equipo editorial]

1 Present address: Department of Ecology and Evolutionary Biology, Princeton University, Guyot Hall, Washington Rd., Princeton, NJ 08544 U.S.A.; Email address: [email protected]

99 100 MARKHAM AND WATTS VOL. 42, NO.2

Prey availability is a key determinant in the distri- METHODS bution (Dzus and Gerrard 1993), density (Gerrard We monitored 18 Bald Eagle nests in the lower et al. 1983), and success (Hansen 1987, Dykstra et al. Chesapeake Bay #3 km inland from the shorelines 1998, Gill and Elliot 2003) of breeding Bald Eagles of the James, York, and Rappahannock rivers during (Haliaeetus leucocephalus). However, directly assessing the 2002 (N 5 8) and 2003 (N 5 10) breeding prey availability in natural systems is difficult due to seasons (Fig. 1). Three areas were recognized along the broad geographic range and flexible foraging the estuarine salinity gradients of these tributaries: habits of Bald Eagles (Gende et al. 1997). As a result, tidal-fresh (0.0–0.5 ppt salinity), oligohaline (0.5– few investigations have examined variation in prey 5.0 ppt), and mesohaline (5.0–18.0 ppt; Chesapeake use within a breeding population and little has been Bay Program Monitoring Subcommittee Data Analy- reported about how diet varies along environmental sis Work Group unpubl. data). We limited nest gradients that determine prey distribution (but see selection to tidal-fresh and mesohaline reaches to Watson et al. 1991, Dzus and Gerrard 1993, Jackman (1) examine extremes in salinity effects within river et al. 2007). systems and (2) because Watts et al. (2006) docu- In a recent investigation of population parame- mented significant differences in breeding density ters for Bald Eagles in the lower Chesapeake Bay, between these salinity zones. Watts et al. (2006) found that shoreline areas sur- To place video-recording equipment, we selected rounding low saline waters support higher breeding nests that had a history of reliable production and densities, have higher reproductive rates, collective- regularly experienced human interaction (with the ly produce a greater number of young, and have reasoning that breeding pairs habituated to human experienced faster rates of population recovery presence were less likely to be disrupted by research- compared to shoreline areas surrounding higher er activity). Using these criteria, we selected nine saline waters. This implies tidal-fresh reaches repre- nests in each salinity zone. One nest in the mesoha- sent core breeding areas for Bald Eagles, yet the line reach was used in both years of this study; be- ecological advantages of these regions are un- cause nests were considered individual samples in known. One possible explanation for the observed all analyses, samples sizes presented reflect the pattern is that the influence of salinity on eagle unique pairing of nest location and year. breeding density is mediated through prey availabil- Data collection at each nest was determined by ity (Watts et al. 2006). In the Chesapeake Bay, Bald nestling age, as estimated during aerial surveys Eagles forage primarily on fish during their breed- and later confirmed by visual inspection of chicks. ing season (Wallin 1982, Mersmann 1989). Fish We divided the nesting cycle into three phases rel- communities are not uniform throughout the estu- ative to the expected period of maximum growth in arine ecosystem and salinity is one of the key factors developing eaglets: before (0–14 d), during (15– known to influence the abundance and distribution 45 d), and after (46 d–fledging) anticipated maxi- of fish species in Chesapeake Bay waters (Murdy et mum growth (Bortolotti 1984). Recording effort at al. 1997, Jung 2002). This, along with data indicat- all nests was focused primarily on the maximum ing that Bald Eagle breeding pairs typically forage growth phase. By this age, eaglets are endothermic within home ranges close to their nest site (,3 km; (14.7 d; Bortolotti 1984) and nest trees can be Buehler et al. 1991), suggests that eagles nesting in climbed for camera installation with minimal risk different salinity zones encounter different suites of associated with exposing chicks to ambient temper- prey species or experience different food resource atures. Further, nestlings in this phase experience levels. the fastest rate of growth (Ricklefs 1967) and, ac- Here, we investigate the influence of salinity on cordingly, provisioning rates have the greatest im- the diet composition of Bald Eagles in the lower pact on overall growth patterns (Bortolotti 1989). Chesapeake Bay to examine how breeding pairs re- Monitoring for two nests began at hatching, and spond to potential differences in prey communities thus, also included the pre-maximum growth phase and availability between salinity zones. Specifically, (cameras were installed prior to egg-laying); moni- we (1) describe the diet of breeding Bald Eagles in toring of 10 nests continued through fledging and the Chesapeake Bay region, (2) evaluate the extent therefore included the post-maximum phase. For to which salinity contributes to spatial variation in nests with multiple young, we used the hatch date diet, and (3) investigate patterns of variation in diet of the oldest nestling when assigning brood ages for between the two years of this study. data analysis. Hatch date was either (1) determined JUNE 2008 BALD EAGLE DIET 101

Figure 1. Locations of nests used (2002–2003) in the lower Chesapeake Bay study area. Nests are distinguished between those in tidal-fresh (N) and mesohaline (&) salinity zones. from video-monitoring or (2) estimated based upon 250 m from the nest to reduce disturbance during visual inspection of nestlings when they were be- maintenance activities. Recording of nest activity tween 15–20 d. Nests in both salinity reaches (tidal- was focused on the morning hours (beginning fresh and mesohaline) included broods ranging in 1 hr after sunrise) to include the expected peak size from 1–3 chicks. Brood size did not vary with period of chick provisioning ( Jaffe 1980, Wallin salinity for nests included in this study. 1982, Mersmann 1989). Recording bouts typically Video-monitoring. The video system consisted of lasted 8 hr, the duration of standard T-160 VHS a waterproof, bullet security camera wired to a VHS videotapes. Each nest was monitored in this fashion videocassette recorder. We used both color and approximately 4 d/wk with effort taken to maintain black-and-white cameras. Video cameras were equal sampling between salinity zones within each mounted to the nest tree approximately 1 m above study year. the nest so that the entire nest surface was in view. During video review, we identified prey items to Cameras were wired to a videorecorder and a deep the lowest taxonomic level possible and recorded cycle, 12-volt marine battery. The videorecorder and date, delivery time, and prey size. We grouped prey battery were placed in waterproof containers and into four taxonomic classes (fish, birds, mammals, positioned at a remote location approximately and reptiles) and developed methods for determin- 102 MARKHAM AND WATTS VOL. 42, NO.2 ing size and estimating biomass that reflected the ity zones. We used a one-way analysis of variance morphological characteristics of each taxa. Similar (ANOVA) with year (two classes: 2002 and 2003) methods were used to assess size and biomass of as a factor and individual nests as samples to evalu- unidentified prey items. For all deliveries that were ate the influence of year on overall species diversity. not whole, we estimated the intact proportion in Only provisioning data collected during the period 10% intervals. of maximum growth was used in further salinity We estimated the size of fish by visual comparison zone comparisons. to adult bill length in half-bill-length increments. We evaluated the influence of salinity and year on We then converted observations to mm using an diet using two-way ANOVAs with salinity (two clas- estimate of mean bill length for eagles from the ses: tidal-fresh and mesohaline) and year (2002 and Chesapeake Bay population. The estimate (63.6 6 2003) as factors and individual nests as samples. We 3.40 mm; mean 6 SD) used was derived from 26 tested a series of response variables including diet adult and subadult specimens (11 females, 12 breadth and equitability, mean fish length and bio- males, and three unsexed) housed in the bird col- mass, the importance of Ictaluridae and Clupeidae lections of the Smithsonian National Museum of fish, and the importance of other fish species (not Natural History and Virginia Tech (B. Watts and Ictaluridae and Clupeidae) collectively. At each A. Markham unpubl. data). Fish lengths were then nest, we computed species diversity (Simpson’s D: converted to biomass using species-specific length- Simpson 1949) to estimate diet breadth and equita- mass relationships (Appendix). We used values bility to estimate evenness. Equitability J was calcu- from closely related species if no length-mass con- lated as J 5 H log (S) where H is Shannon-Wiener versions were available. For taxa with species mem- index and S is the number of species in the diet bers that were indistinguishable on videotape (e.g., (Pielou 1966). For fish species and groups, we used Ictaluridae, Clupeidae, Lepomis spp., and Pomoxis biomass delivery rate (g/h) and percentage of total spp.), biomass conversions were based on represen- fish biomass delivered as estimates of importance tative species. Biomass calculations for unidentified relative to the overall eagle diet. Because recording fish were estimated from a weighted average of all effort varied between nests, we used a Michaelis- identified fish. Menton function to determine if diet breadth had Birds were identified and assigned species-specific reached an asymptote within the recordings for masses from Dunning (1992). Mammals were classi- each nest (Miller and Wiegert 1989). Two nests that fied as either juvenile or adult depending on size did not have adequate recording effort for diet to (small or large) and assigned masses specific to re- reach an asymptote were not included in salinity gional specimens and appropriate for correct age/ comparisons. All statistical computations were per- size category (Nowak and Paradiso 1983). We also formed using the software package STATISTICA v.5 estimated turtle size (carapace length) in relation to (StatSoft, Inc. 1995) with an alpha value for statisti- adult bill length. Whole weights for biomass conver- cal significance set to 0.05. sions were estimated using species-specific allome- tric relationships between carapace length and mass RESULTS derived from regional datasets ( J. Mitchell unpubl. Brood Size. Brood size for all nests observed for data). To be conservative in our calculations, we prey deliveries varied between one and three young estimated the biomass of all unidentified prey items with a mean of 1.8 6 0.73 (SD). Brood size was in relation to the approximate size (and associated higher in nests associated with tidal-fresh reaches mass) of adult mammals. compared to nests associated with mesohaline Statistical Analyses. We summarized all prey deliv- reaches (2.0 6 0.71 and 1.6 6 0.74, respectively) eries by presenting descriptive statistics on overall but this difference was not statistically distinguish- diet composition and prey size for all taxa to pro- able (F1,16 5 1.13, P 5 0.30). vide an overview of eagle diet within the study area Overall Diet. Of the 765 prey deliveries recorded during the brood-rearing period. We evaluated the on 4098 hr (18 nests) of videotape, we identified possible influence of year and salinity on the use of 730 (95.4%) prey items representing at least 12 spe- broad taxonomic classes using a chi-square where cies of fishes, three species of birds, four species of the expected distribution of prey between the clas- mammals, and four species of reptiles (Table 1). ses was derived from 2002 for comparisons between Some fish species were indistinguishable on video- years and tidal-fresh for comparisons between salin- tape and only family or genus level identification JUNE 2008 BALD EAGLE DIET 103

Table 1. Diet of Bald Eagles nesting in the lower Chesapeake Bay based on video-observations of prey delivered to nests during the 2002–2003 breeding seasons.

INDIVIDUALS (N)BIOMASS (kg)

SPECIES TOTAL %TOTAL % Fish (Osteichthyes) American eel (Anguilla rostrata) 17 2.2 11.75 3.8 Atlantic croaker (Micropogonias undulatus) 45 5.9 20.55 6.6 Black crappie and white crappie (Pomoxis spp.) 10 1.3 2.43 0.8 Bluefish (Pomatomus saltatrix) 1 0.1 0.11 0.0 Clupeidae 284 27.1 111.05 26.4 Ictaluridae 234 30.6 95.30 30.7 Largemouth bass (Micropterus salmoides) 10 1.3 7.45 2.4 Lepomis spp. 10 1.3 2.45 0.8 Spot (Leiostomus xanthurus) 2 0.3 0.89 0.3 Striped bass (Morone saxatilis) 5 0.7 6.63 2.1 Summer flounder (Paralichthys dentatus) 1 0.1 0.46 0.1 Yellow perch (Perca flavescens) 1 0.1 0.34 0.1 Unidentified 75 9.8 19.14 6.2 Fish Subtotal 695 90.8 278.56 89.8 Birds (Aves)a 7 0.9 4.38 1.4 Mammals (Mammalia)b 20 2.6 12.65 4.1 Reptiles (Reptilia)c 8 1 2.42 0.8 Unidentified Prey 35 4.6 12.17 3.9 GRAND TOTAL 765 100.0 310.18 100.0 a Identified bird species: Double-crested Cormorant (Phalacrocorax auritus), Mallard (Anas platyrhynchos), and Rock Pigeon (Columba livia). b Identified mammal species: common muskrat (Ondatra zibethicus), eastern cottontail (Sylvilagus floridanus), eastern gray squirrel (Sciurus carolinensis), and woodchuck (Marmota monax). c Identified reptile species: common musk turtle (Sternotherus odoratus), eastern mud turtle (Kinosternon subrubrum), eastern painted turtle (Chrysemys picta picta), and snapping turtle (Chelydra serpentina). was possible. Specifically, we were not able to constituted 94.1% of the biomass delivered, birds discriminate between species within two families (Ic- 1.0%, mammals 4.1%, and reptiles 0.9%. Ictaluridae taluridae and Clupeidae). Probable species repre- and Clupeidae were overwhelmingly the most com- sented were: channel catfish (Ictalurus punctatus), mon prey groups by frequency and biomass. blue catfish (I. furcatus), and white catfish (Ameiurus We were able to calculate biomass for 750 prey catus) in the Ictaluridae family; and alewife (Alosa items (98.9% of observed deliveries). Including pseudoharengus), American shad (A. sapidissima), items of both partial and intact prey status, biomass blueback herring (A. aestivalis), hickory shad (A. of individual deliveries ranged from 1.3–2391.7 g mediocris), and gizzard shad (Dorosoma cepedianum) with a mean of 414.5 g (SD 5 291.5). Of the 620 in the Clupeidae. In addition, we were unable identified fish, we were able to determine intact to discriminate between species within two gen- status and estimate total length for 473 items; bio- era (Lepomis spp. and Pomoxis spp). Probable species mass of intact fish ranged from 22.8–2391.7 g with a represented were: bluegill (Lepomis macrochirus), mean of 466.5 g (SD 5 268.70). pumpkinseed (L. gibbosus), and redbreast sunfish Neither the importance of each taxonomic class (L. auritus) in the Lepomis genus; and black crappie (by percent biomass of all prey identified to class) in (Pomoxis nigromaculatus) and white crappie (P. annu- overall diet composition throughout the breeding 2 laris)inPomoxis. season (x 3 5 0.31, P 5 0.96) nor species diversity By frequency of occurrence, fish constituted (F1,14 5 0.03, P 5 0.87) varied significantly between 96.0% of the identified prey, birds 0.5%, mammals the two years of this study. Ictaluridae and Clupei- 2.3%, and reptiles 1.2%. By delivered biomass, fish dae were dominant in the diet for both study years, 104 MARKHAM AND WATTS VOL. 42, NO.2

Figure 2. Influence of salinity and year on biomass delivery rate of Ictaluridae at Bald Eagle nests in the lower Chesapeake Bay during the 2002–2003 breeding seasons. Only deliveries made during the expected period of nestling maximum growth (15–45 d) were included in analysis. with no significant between-year variation in com- The combined dietary proportion of Ictaluridae 2 bined use (x 1 5 0.20, P 5 0.65). and Clupeidae expressed as percent biomass of de- Salinity and Yearly Comparisons. We observed livered fish biomass did not vary between salinity 541 prey deliveries during 2176 hr of videorecord- zones (F1,12 5 0.89, P 5 0.36) or years (F1,12 5 ing at the 16 nests with recording coverage ade- 4.58, P 5 0.054) and no interaction was detected quate for statistical analysis. The proportion of diet (F1,12 5 0.88, P 5 0.37). The proportion of Ictalur- represented by each taxonomic class (by percent idae biomass in the diet was lower in 2002 com- biomass of all prey identified to class) did not vary pared to 2003 (F1,12 5 4.59, P 5 0.05; Fig. 2); the 2 significantly between salinity zones (x 3 5 4.5, P 5 proportion of Clupeidae biomass showed the re- 2 0.21) or years (x 3 5 8.8, P 5 0.07). Fish dominated verse trend (F1,12 5 6.04, P 5 0.03; Fig. 3). There diet composition in both salinity zones and during was no significant effect of salinity on the dietary both years. proportion of Ictaluridae (F1,12 5 0.34, P 5 0.57) The length (F1,469 5 7.50, P 5 0.006) and biomass or Clupeidae biomass (F1,12 5 0.61, P 5 0.45), al- (F1,469 5 7.56, P 5 0.006) of delivered fish were though there was a trend toward a higher biomass higher in the mesohaline compared to tidal-fresh proportion of each family in tidal-fresh reaches reaches. Mean fish length was 40.0 (67.48) cm compared to mesohaline zones in 2002. No signifi- and 42.4 (610.39) cm in tidal-fresh and mesohaline cant interaction was observed between salinity and zones, respectively. Fish mass was 434.9 (6235.64) g year for either fish family (F1,12 , 0.34, P 5 0.57). and 509.6 (6303.43) g for tidal-fresh and mesoha- We also examined the prevalence of Ictaluridae and line zones, respectively. We observed no significant Clupeidae in the diet by considering the rates of between year differences in length (F1,469 5 0.18, biomass delivery for each family (Figs. 2 and 3). P 5 0.67) or biomass (F1,469 5 0.15, P 5 0.70). No statistically significant results were observed be- Species diversity (Simpson’s D) was higher in the tween salinity zones (F1,12 , 0.36, P 5 0.56) or years mesohaline compared to tidal-fresh reaches (F1,12 5 (F1,12 , 2.73, P 5 0.12) and no interaction was de- 4.65, P 5 0.052). However, there was no significant tected (F1,12 , 0.28, P 5 0.61). Ictaluridae and Clu- difference between years in species diversity (F1,12 5 peidae showed reverse patterns between years with 0.22, P 5 0.65), though there was a trend toward Ictaluridae delivery rates being higher in 2003 com- higher diversity in 2002. Equitability was not differ- pared to 2002 and Clupeidae delivery rates having ent between salinity zones (F1,12 5 3.07, P 5 0.011) the opposite pattern. However, in 2002 both prey or years (F1,12 5 0.49, P . 0.50). groups exhibited similar spatial patterns with deliv- JUNE 2008 BALD EAGLE DIET 105

Figure 3. Influence of salinity and year on biomass delivery rate of Clupeidae at Bald Eagle nests in the lower Chesa- peake Bay during the 2002–2003 breeding seasons. Only deliveries made during the expected period of nestling maximum growth (15–45 d) were included in analysis. ery rates being higher in tidal-fresh compared to (e.g., Watson et al. 1991, Ewins and Andress 1995). mesohaline reaches. Finally, no statistically signifi- In addition, our findings indicated a greater pro- cant results were observed in the rates of biomass portion of small and soft-bodied fish relative to delivery for other fish species (i.e., excluding Icta- results from previous research in which diet was luridae and Clupeidae) between salinity zones (F1,12 assessed by analyzing prey remains in or under nests 5 0.26, P 5 0.62) or years (F1,12 5 1.11, P 5 0.31) (Cline and Clark 1981, Haines 1986). This is likely and no interaction was detected (F1,12 5 0.49, P 5 due to the eagle’s ability to digest the fine bones 0.50). associated with small mammals and many fish species (Imler and Kalmbach 1955, Duke et al. 1975). DISCUSSION Thus, the differences between past and current Overall Diet. Important prey identified in this findings likely reflect the improved accuracy of diet investigation were similar to those noted in earlier assessment using a video-sampling approach (Mers- studies of Bald Eagle diet in the Chesapeake Bay mann et al. 1992). (e.g., Wallin 1982, Mersmann 1989); however, dif- Ictaluridae and Clupeidae prey dominated the ferences in the relative use of key species and prey diet composition of breeding pairs in this study. taxa were apparent. We suggest that these discrep- The importance of catfish in the diet of Bald Eagles ancies were largely the result of (1) variation in the has been reported in numerous other foraging stud- foraging ecology of breeding versus nonbreeding ies throughout the species’ range (e.g., McEwan eagles and (2) differences in the field techniques and Hirth 1980, Cash et al. 1985, Dugoni et al. used to assess diet. For example, previous research 1986, Mabie et al. 1995). Bald Eagle predation on focused on diet composition during the breeding both nonmigratory (e.g., Mersmann 1989, Mabie et season was consistent with our results that fish con- al. 1995) and migratory (e.g., Fitzner and Hanson stitute the overwhelming majority of items delivered 1979, Spencer et al. 1991, Hunt et al. 1992, McClel- to nests (.98%; Wallin 1982). Outside the breeding land et al. 1994) Clupeidae has also been previously season, eagles in the Chesapeake Bay ecosystem re- documented. In the areas where research on Bald lied more heavily on other prey taxa such as birds Eagles and anadromous fish interactions has been and mammals (Mersmann 1989). This change in conducted, spawning runs do not coincide with the diet has been correlated with seasonal shifts in prey eagle breeding season. Two notable exceptions were abundance and the eagle’s ability to forage oppor- described by Gerrard et al. (1975) and Gende et al. tunistically on temporally abundant food resources (1997), who documented higher productivity in ea- 106 MARKHAM AND WATTS VOL. 42, NO.2 gle pairs nesting close to spawning grounds com- dance occurring at upriver sampling stations where pared with pairs nesting farther away. Thus, a find- river salinities are highly diluted (VIMS unpubl. da- ing of particular interest here is the novelty of ta). This finding was consistent with the trend we breeding eagles utilizing anadromous species within observed of higher catfish delivery rates and higher the Chesapeake Bay ecosystem. dietary proportion of catfish in lower salinity zones. Salinity and Yearly Comparisons. Previous re- Fishes in the Clupeidae family are predominantly search indicated that Bald Eagles seemed to alter anadromous species that migrate from open bay or prey size selection based upon energetic require- ocean waters to spawn in the freshwater portions of ments. Jenkins and Jackman (1994) concluded that creeks and rivers. Within the Chesapeake Bay sys- differences observed between breeding and non- tem, spawning runs are triggered by favorable water breeding eagles with regard to mean prey size sup- temperatures and typically extend from February ported optimal foraging models. In our study, adults through June (alewife, Munroe 2000; hickory shad, on average delivered significantly larger fish (with J. David, J. Miller, and W. Wilson unpubl. data). regard to fish length and biomass) to nests in the This time window is temporally synchronous with mesohaline compared to tidal-fresh salinity zones. the nesting season of Bald Eagles, particularly coin- This suggested more energetically favorable condi- ciding with the period in which most nestlings in tions for foraging parents in higher salinity waters if Virginia experience maximum growth, and thus, adults in both salinity zones expended similar have the greatest energetic demands (April–mid- amounts of effort per capture. Further evaluation May). As in other predator species (Willson and of this assumption is needed to determine the extent Halupka 1995), Bald Eagles may have responded to which available prey sizes may influence patterns to the predictable and concentrated influx of ener- in nesting density and reproductive rates with salini- gy-rich food resources associated with these annual ty. In particular, an investigation of the energetic spawning runs. consequences of prey size variation during the breed- Unlike the anadromous Clupeidae, gizzard shad ing season (e.g., increased rates of nestling provision- are year-round residents of the Chesapeake Bay eco- ing and growth) may add additional insight into un- system (Murdy et al. 1997). Gizzard shad are widely derstanding nest density patterns. distributed through the bay waters with high abun- Our results showed that the diet composition of dances in the freshwater portions of the tributaries breeding Bald Eagles did not vary spatially along the up to salinities as high as 22 ppt (Murdy et al. 1997). salinity gradient of the lower Chesapeake Bay tribu- Because we were unable to distinguish different Clu- taries. In both tidal-fresh and mesohaline salinity peidae species, the relative importance of gizzard zones, the diet was dominated by Ictaluridae and shad compared to anadromous Alosa spp. remains Clupeidae species. Additional fish species observed uncertain. tended to occur in such small numbers that spatial Record wet conditions and lower than average differences in distribution were not detectable. This temperature were observed in 2003 throughout suggested that changes in salinity between tidal- the southeast United States, including Virginia fresh and mesohaline zones did not impose a signif- (Gleason et al. 2004). We consider these differences icant constraint on diet composition as originally in rainfall and temperature between study years as hypothesized. the most likely factor influencing the observed The broad-scale use of Ictaluridae and Clupeidae annual variation in diet composition. Increased may be interpreted through taxa-specific distribu- rainfall results in a rise in freshwater input, which tion patterns and life history characteristics. Al- consequently affects water flow, temperature, and though catfish belonging to the family Ictaluridae turbidity in the tributary systems. These factors, in generally prefer low-saline waters, they are capable turn, negatively affect migration runs of anadro- of utilizing a broad salinity range (Lippson et al. mous species ( Jung 2002) and could be responsible 1979, Dames et al. 1989, Jenkins and Burkhead for the decreased relative importance of Clupeidae 1994). Therefore, catfish seem to be available to in the diet in 2003 compared to 2002. In addition, eagles nesting in both tidal-fresh and mesohaline increased freshwater input influences salinity salinity zones. However, trawl surveys conducted by boundaries, effectively extending the freshwater the Virginia Institute of Marine Science (VIMS) in portions of the tributaries. Reduced salinities may the lower 35 km of the James, York, and Rappahan- have expanded the foraging range of Ictaluridae nock rivers reported greatest overall catfish abun- species (Sauls et al. 1998), and thus, also explain JUNE 2008 BALD EAGLE DIET 107 the increased relative importance of catfish use in DZUS, E.H. AND J.M. GERRARD. 1993. Factors influencing the mesohaline reaches in 2003. However, this Bald Eagle densities in northcentral Saskatchewan. J. alone does not explain the increased use of catfish Wildl. Manage. 57:771–778. in the tidal-fresh reaches in 2003. EWINS, P.J. AND R.A. ANDRESS. 1995. The diet of Bald Ea- gles, Haliaeetus leucocephalus, wintering in the lower ACKNOWLEDGMENTS Great Lakes basin. Can. Field-Nat. 109:418–425. FESSLER, F.R. 1949. A survey of fish populations in small We owe special thanks to the numerous volunteers and ponds by two methods of analysis. M.S. thesis, Iowa landowners whose time and generosity made this study possible. We also thank Drs. H. Austin, M. Byrd, P. Heideman, State College, Ames, IA U.S.A. and three anonymous reviewers for their valuable com- FITZNER, R.E. AND W.C. HANSON. 1979. A congregation of ments on earlier drafts of this manuscript. B. Paxton offered wintering Bald Eagles. Condor 81:311–313. expertise in the field and created the study map. J. Mitchell FORTIN,R.AND E. MAGNIN. 1972. Croissance en longueur et shared morphometric data on turtle specimens essential in en poids des perchaudes Perca flavescens de la Grande length-weight regression calculations. The Smithsonian Mu- Anse de l’ile Perrot au lac Saint-Louis. J. Fish. Res. Board seum of Natural History and Virginia Tech allowed access to Can. 29:517–523. the Bald Eagle skins in their bird collections. This study was financially supported by the Center for Conservation Biolo- GENDE, S.M., M.F. WILLSON, AND M. JACOBSEN. 1997. Repro- gy and grants from the College of William and Mary, the ductive success of Bald Eagles (Haliaeetus leucocephalus) Mary and Daniel Loughran Foundation, the Northern Neck and its association with habitat or landscape features Audubon Society, the Virginia Society of Ornithology, and and weather in southeast Alaska. Can. J. Zool. 75:1595– the Williamsburg Bird Club. 1604. GERRARD,J.M.,P.N.GERRARD,G.R.BORTOLOTTI, AND LITERATURE CITED D.W.A. WHITFIELD. 1983. A 14-year study of Bald Eagle BORTOLOTTI, G.R. 1984. Physical development of nestling reproduction on Besnard Lake, Saskatchewan. Pages Bald Eagles with emphasis on the timing of growth 47–58 in D.M. Bird [ED.], Biology and management events. Wilson Bull. 96:524–532. of Bald Eagles and Ospreys. Harpell Press, Ste. Anne ———. 1989. Factors influencing the growth of Bald Eagles de Bellevue, Quebec, Canada. in north central Saskatchewan. Can. J. Zool. 67:606–611. ———, ———, W.J. MAHER, AND D.W.A. WHITFIELD. 1975. BUEHLER, D.A., T.J. MERSMANN,J.D.FRASER, AND J.K.D. Factors influencing nest site selection of Bald Eagles in SEEGAR. 1991. Differences in distribution of breeding, northern Saskatchewan and Manitoba. Blue Jay 33:169–176. nonbreeding, and migrant Bald Eagles on the north- GILL, C.E. AND J.E. ELLIOTT. 2003. Influence of food supply ern Chesapeake Bay. Condor 93:399–408. and chlorinated hydrocarbon contaminants on breed- CASH, R.J., P.J. AUSTIN-SMITH,D.BANKS,D.HARRIS, AND P.C. ing success of Bald Eagles. Ecotoxicology 12:95–111. SMITH. 1985. Food remains from Bald Eagle nest sites GLEASON, K.L., A.M. WAPLE,T.EICHLER, AND G.D. BELL. on Cape Breton Island, Nova Scotia. J. Wildl. Manage. 2004. Regional climate, North America, United States 49:223–225. of America. Pages S32–S33 in D.H. Levinson and A.M. CLINE, K.W. AND W.S. CLARK. 1981. Chesapeake Bay Bald Waple [EDS.], State of the climate in 2003. Bull. Am. Eagle banding project: 1981 report and five-year sum- Meteorol. Soc. 85(6). mary. Raptor Information Center, Natl. Wildl. Fed., HAIMOVICI,M.AND G. VELASCO. 2000. Length-weight rela- Washington, DC U.S.A. tionships of marine fishes from southern Brazil. Naga DAMES, H., T. COON, AND J. ROBINSON. 1989. Movements of ICLARM Quarterly 23:19–23. channel and flathead catfish between the Missouri Riv- HAINES, S.L. 1986. The feeding, roosting, and perching er and a tributary, Perch Creek. Trans. Am. Fish. Soc. behavior of the Bald Eagle (Haliaeetus leucocephalus) 188:670–679. of Mason Neck, Virginia, with special reference to the DAWSON, C.E. 1965. Length-weight relationships of some development of Mason Neck State Park. M.S. thesis, Gulf of Mexico fishes. Trans. Am. Fish. Soc. 94:279–280. George Mason Univ., Fairfax, VA U.S.A. DUGONI, J.A., P.J. ZWANK, AND G.C. FURMAN. 1986. Foods of HANSEN, A.J. 1987. Regulation of Bald Eagle reproductive nesting Bald Eagles in Louisiana. Raptor Res. 20:124–127. rates in southeast Alaska. Ecology 68:1387–1392. DUKE, G.E., A.A. JEGERS,G.LOFF, AND O.A. EVANSON. 1975. HANSEN, R.A. AND A.G. EVERSOLE. 1984. Age, growth and Gastric digestion in some raptors. Comp. Biochem. Phys- sex ratio of American eels in brackish-water portions of a iol. 50A:646–656. South Carolina river. Trans. Am. Fish. Soc. 113:744–749. DUNNING, J.B. 1992. Handbook of avian body weights. HENDERSON, E.M. 1979. Summer flounder (Paralichthys den- C.R.C. Press, Orlando, FL U.S.A. tatus) in the northwest Atlantic. National Marine Fish- DYKSTRA, C.R., M.W. MEYER, D.K. WARNKE, W.H. KARASOV, eries Service, Northeast Fisheries Center, Woods Hole D.E. ANDERSEN,W.W.BOWERMAN,IV,AND J.P. GIESY. Laboratory Reference Number 79–31. 1998. Low reproductive rates of Lake Superior Bald HUNT, W.G., B.S. JOHNSON, AND R.E. JACKMAN. 1992. Carry- Eagles: low food delivery rates or environmental con- ing capacity for Bald Eagles wintering along a north- taminants? J. Gt Lakes Res. 24:32–44. western river. J. Raptor Res. 26:49–60. 108 MARKHAM AND WATTS VOL. 42, NO.2

IMLER, R.H. AND E.R. KALMBACH. 1955. The Bald Eagle and MUNROE, T.A. 2000. An overview of the biology, ecology, its economic status. U.S. Fish and Wildl. Service Circu- and fisheries of the Clupeoid fishes occurring in the lar No. 30. Gulf of Maine. Northeast Fisheries Science Center Ref. JACKMAN, R.E., W.G. HUNT, AND N.L. HUTCHINS. 2007. Bald Doc. 00-02, Woods Hole, MA U.S.A. Eagle foraging and reservoir management in northern MURDY, E.O., R.S. BIRDSONG, AND J.A. MUSICK. 1997. Fishes California. J. Raptor Res. 41:202–211. of the Chesapeake Bay. Smithsonian Institution Press, JAFFE, N.B. 1980. Nest site selection and foraging behavior Washington, DC U.S.A. of the Bald Eagle (Haliaeetus leucocephalus) in Virginia. NOWAK, R.M. AND J.L. PARADISO. 1983. Walker’s mammals M.A. thesis, College of William and Mary, Williams- of the world, Fourth Ed. Johns Hopkins Univ. Press, burg, VA U.S.A. Baltimore, MD U.S.A. JENKINS, J.M. AND R.E. JACKMAN. 1994. Field experiments in PIELOU, E.C. 1966. The measurement of diversity in differ- prey selection by resident Bald Eagles in the breeding ent types of biological collections. J. Theor. Biol. and nonbreeding season. J. Field Ornithol. 65:441–446. 13:131–144. JENKINS, R.E. AND N.M. BURKHEAD. 1994. Freshwater fishes RICKLEFS, R.E. 1967. A graphical model for fitting equa- of Virginia. American Fisheries Society, Bethesda, MD tions to growth curves. Ecology 48:978–983. U.S.A. SAULS, B., D. DOWLING,J.ODENKIRK, AND E. COSBY. 1998. JUNG, S. 2002. Fish community structure and the spatial Catfish population in Chesapeake Bay. Chesapeake Bay and temporal variability in recruitment and biomass Program, Environmental Protection Agency, Annapo- production in Chesapeake Bay. Ph.D. dissertation, lis, MD U.S.A. Univ. of Maryland, College Park, MD U.S.A. SHIELDS, J.T. 1955. Report of fisheries investigations during LIPPSON, A.J., M.S. HAIRE, A.F. HOLLAND,F.JACOBS,J.JEN- the second year of impoundment of Fort Randall Res- SEN, R.L. MORAN-JOHNSON, T.T. POLGAR, AND W.A. RICH- ervoir, South Dakota, 1954. Progress Report 55-4. KUS. 1979. Environmental atlas of the Potomac estuary. South Dakota Dept. Game, Fish and Parks, Pierre, SD Martin Marietta Corporation for MD Dept. Natural Re- U.S.A. sources, Baltimore, MD U.S.A. SIMPSON, E.H. 1949. Measurement of diversity. Nature MABIE, D.W., M.T. MERENDINO, AND D.H. REID. 1995. Prey 163:688. of nesting Bald Eagles in Texas. J. Raptor Res. 29:10–14. SPENCER, C.N., B.R. MCCLELLAND, AND J.A. STANFORD. 1991. MANSUETI, R. 1961. Age, growth and movements of the Shrimp stocking, salmon collapse, and eagle displace- striped bass, Roccus saxatilis, taken in size selective gear. ment. Bioscience 41:14–21. Chesapeake Sci. 2:9–36. STATSOFT,INC. 1995. Statistica for Windows version 5.0. MCCLELLAND,B.R.,L.S.YOUNG,P.T.MCCLELLAND,J.G. Tulsa, OK U.S.A. CRENSHAW, H.L. ALLEN, AND D.S. SHEA. 1994. Migration WALLIN, D.O. 1982. The influence of environmental con- ecology of Bald Eagles from autumn concentrations in ditions on the breeding behavior of the Bald Eagle Glacier National Park, Montana. Wildl. Monogr. (Haliaeetus leucocephalus) in Virginia. M.A. thesis, Col- 125:1–61. lege of William and Mary, Williamsburg, VA U.S.A. MCEWAN, L.C. AND D.H. HIRTH. 1980. Food habits of the WATSON, J.W., M.G. GARRETT, AND R.G. ANTHONY. 1991. Bald Eagle in north-central Florida. Condor 82:229–231. Foraging ecology of Bald Eagles in the Columbia River MERSMANN, T.J. 1989. Foraging ecology of Bald Eagles on estuary. J. Wildl. Manage. 55:492–499. the northern Chesapeake Bay with an examination of WATTS, B.D., A.C. MARKHAM, AND M.A. BYRD. 2006. Salinity techniques used in the study of Bald Eagle food habits. and population parameters of Bald Eagles (Haliaeetus M.S. thesis, Virginia Polytechnic Institute and State leucocephalus) in the lower Chesapeake Bay. Auk Univ., Blacksburg, VA U.S.A. 123:393–404. ———,D.A.BUEHLER,J.D.FRASER, AND J.K.D. SEEGAR. WILK,S.J.,W.W.MORSE, AND D.E. RALPH. 1978. Length- 1992. Assessing bias in studies of Bald Eagle food hab- weight relationships of fishes collected in the New York its. J. Wildl. Manage. 56:73–78. Bight. Bull. NJ Acad. Sci. 23:58–64. MILLER, R.J. AND R.G. WIEGERT. 1989. Documenting com- WILLSON, M.F. AND K.C. HALUPKA. 1995. Anadromous fish pleteness, species-area relationships, and the species- as keystone species in vertebrate communities. Conserv. abundance distribution of a regional flora. Ecology Biol. 9:489–497. 70:16–22. MUNCY, R.J. 1959. Age and growth of channel catfish from the Des Moines River, Boone County, Iowa, 1955 and 1956. Iowa J. Sci. 34:127–137. ———. 1960. A study of the comparative efficiency be- Received 1 November 2005; accepted 18 January 2008 tween nylon and linen gillnets. Chesapeake Sci. 1:96–102. Associate Editor: James C. Bednarz JUNE 2008 BALD EAGLE DIET 109

Appendix. Biomass conversions used for fish and turtle species identified in the diet of Bald Eagles nesting in the lower Chesapeake Bay during 2002–2003 breeding seasons. In conversion equations, mass (M) in grams and length (L) in centimeters.

SPECIES BIOMASS CONVERSION REFERENCE American eel (Anguilla rostrata)M5 0.00166 3 L3.07 Hansen and Eversole 1984 Atlantic croaker (Micropogonias undulatus)M5 0.0031 3 L3.25 Wilk et al. 1978 Bluefish (Pomatomus saltatrix)M5 0.01356415 3 L2.898996 Haimovici and Velasco 2000 Clupeidae (used American shad, Alosa sapidissima)M5 0.0065 3 L2.959 Muncy 1960 Ictaluridae (used channel catfish, Ictalurus punctatus)M5 0.00397 3 L3.133 Muncy 1959 Largemouth bass (Micropterus salmoides)M5 0.00728 3 L3.113 Fessler 1949 Lepomis spp. (used bluegill, L. macrochirus)M5 0.00698 3 L3.209 Fessler 1949 Pomoxis spp. (used black crappie, P. nigromaculatus)M5 0.0101 3 L3.074 Shields 1955 Spot (Leiostomus xanthurus)M5 0.00921 3 L3.072 Dawson 1965 Striped bass (Morone saxatilis)M5 0.00614 3 L3.153 Mansueti 1961 Summer flounder (Paralichthys dentatus)M5 0.00544 3 L3.117 Henderson 1979 Yellow perch (Perca flavescens)M5 0.00785 3 L3.083 Fortin and Magnin 1972 Common musk turtle (Sternotherus odoratus)M5 0.24400 3 L2.7819 J. Mitchell unpubl. data Eastern mud turtle (Kinosternon subrubrum)M5 0.25300 3 L3.0220 J. Mitchell unpubl. data Eastern painted turtle (Chrysemys picta picta)M5 0.30500 3 L2.6480 J. Mitchell unpubl. data Snapping turtle (Chelydra serpentina)M5 0.22000 3 L3.0220 J. Mitchell unpubl. data